Updated with responses from CNPP, June 24, 2019. The complete letter from CNPP can be viewed here.
by Nina Teicholz
The advisory committee appointed to oversee the science for the next set of Dietary Guidelines, in 2020, held its inaugural meeting last month—with some startling surprises. Most extraordinary was the government’s assertion that at a time when 60% of the population is afflicted with some kind of nutrition-related disease, the Guidelines will continue to be a policy for healthy Americans only.
With such a narrow focus, this policy is on track to do virtually nothing to reverse the epidemics of disease that are causing enormous suffering as well as bankrupting our country. What’s more, this narrow scope could be illegal. A 1990 law states that the Dietary Guidelines must be for the “general public” (Sec. 301)—which is now rightly defined as those of us with nutrition-related diseases.
One committee member took issue with this narrow focus, querying the USDA staff: “But if you excluded such people [with a particular disease condition]…that would not actually be representative of who lives in this country” (Day 1, Morning, 47:00). Exactly.
Dietary Guidelines Only for “Healthy Americans”
It’s a secret of sorts that the Guidelines exclude people with obesity, diabetes, dementia, heart disease, or other diseases tied to nutrition. After all, the Guidelines virtually dictate feeding programs for people who are both sick and well—“captive” populations in settings such as schools, hospitals, nursing homes and more. The Guidelines are also downloaded as the virtual “gold standard” by medical doctors, nutritionists, dieticians, and nurses alike.
Thus, if you are suffering from obesity, diabetes, dementia, high blood pressure or more, professionals will hand you the government’s one-size-fits-all diet, designed exclusively for healthy people.
This is the equivalent of doling out size 2 clothing to everyone, when more than half the population now wears XXL. It’s not only inappropriate but simply a bad fit--and bad public policy.
Why is the government’s diet not appropriate for those of us who are fat and sick? The last two decades of science indicate that many people with nutrition-related diseases typically have a "‘broken’ metabolism which reduces their tolerance of carbohydrates—found most commonly in grains and sugars. While some people can reverse obesity and diabetes by calorie restriction, others find more success restricting only carbohydrates.
Worse, these people are ill-served by the government’s diet that mandates 50-55% of daily calories as carbohydrates. Indeed, when the American Diabetes Association recently re-issued its guidelines, it stated that a low-carbohydrate diet was the most effective approach for glucose control, which is an essential component in the prevention and management of type 2 diabetes.
Despite this, the USDA’s review of the science for the 2020 Guidelines will not consider this new evidence, because many of the relevant studies have been conducted on sick populations--who are deemed outside the scope of the Guidelines’ process.
USDA officials stated at the meeting that they will not include exclusively studies on sick populations, because the Guidelines are only for healthy Americans.(Day 1, Afternoon, 20:16) [UPDATE: CNPP stated that it will look at studies on people with chronic, nutrition-related diseases but not if those studies look exclusively at people with those diseases. Presumably this means that a study would need to include a mixed population of healthy and unhealthy people to be included in CNPP’s reviews. We’ve followed up to confirm. If true, this criteria would be rather bizarre, because many studies are tailored to narrow populations, in order to obtain clearly delineated conclusions, and are therefore not likely to fit CNPP criteria.]
Importance of Scientific Rigor in the Dietary Guideline
A need for greater “scientific rigor” in the Guidelines was identified as a problem by the National Academy of Sciences, Engineering, and Medicine, in a recent report. 
Responding to the Academies’, USDA staffers at the inaugural meeting talked at length about their efforts to update their systematic review processes, but these explanations were short on details. Moreover, there are a number of worrisome signs that the 2020 Guidelines will suffer from many of the same problems that we’ve seen in the past.
Why is this so important? Quite simply, without good science, we will never get trustworthy dietary guidance.
Lack of Scientific Rigor in the 2020 Dietary Guidelines process
USDA’s office in charge of the Guidelines is called the Center for Nutrition Policy and Promotion (CNPP), and its members are clearly professional, well-educated, and hard working. They ran the two-day meeting with humor and a clear effort to maximize transparency—all to be commended. Yet a number of their decisions indicated that this process has only a slim chance of producing rigorous, evidence-based Dietary Guidelines. Here are some of the major points we observed:
Lack of Specific Protocols
As mentioned above, CNPP staff talked about the agency’s process for improving the systematic reviews. Nutritionist Julie Obbagy discussed plans to adopt a highly regarded review system called “GRADE,” which specifies a rigorous set of protocols for selecting/evaluating science. Adopting GRADE would be an excellent step for CNPP, but is it true? GRADE is not listed in the USDA’s formal response to the Academies nor is it on CNPP’s “Methodology” page. [UPDATE on GRADE: CNPP is not working with any external GRADE experts and says that it has instead developed GRADE expertise in-house. CNPP did not offer any materials documenting its use of GRADE; instead, CNPP cited this page as being the relevant methodology that has been “updated…to align closely with GRADE, while still ensuring it meets the unique needs to help inform population-based nutrition guidance.” CNPP’s methodology differs in several important ways from GRADE—we will elaborate in a separate blog.]
It appears that the agency has no protocols for selecting or evaluating scientific studies
In the meeting, CNPP staff seemed to confirm that they do not yet have protocols for their reviews. In response to a question from the committee, Obbagy said, “They’re not[the protocols]—we haven’t made them yet! That’s sort-of the work that you’ll [the committee] be doing.” (Day 1, afternoon, 50:19)
[UPDATE: CNPP confirmed that they do not, as yet, have any established protocols for their systematic reviews (which is odd, given that GRADE, if employed, would provide such protocols). CNPP states that the various subcommittees will be creating their own protocols; “Draft protocols will be posted online here for the public to view and track the progress of each question under review by the Committee.”]
Obbagy says that her office will continue to uses “manual [hand] searches” for scientific papers (Day 1, Afternoon, 23:30). However, because hand searches do not follow any particular protocol, this process opens up large opportunities for bias. In fact, fully 50% of the papers in the 2015 review on saturated fats relied on “hand searches,” as did 1/3 of the papers in the reviews on the USDA “Dietary Patterns.” This is very hard to explain in a world where nearly 100% of scientific papers are now online .
CNPP is Allowing Study Protocols to be Designed in an Ad Hoc Way?
On Day 2, committee members had a brief conversation in which they made up the criteria for neuro-cognitive disease outcomes. It was clear that none of the committee members are specialists in this area, yet after 15 minutes of superficial discussion with input from only a handful of members, the committee Vice Chair decided on the outcomes they would use, stating, “I hate to say it, but I think we’ve done it!” (Day 2, Afternoon, 10:18 to 26:00) [UPDATE: CNPP says that this conversation was not the last word in establishing outcomes, and that the committee will continue to work on this. If this is so, one might wonder why this particular discussion took place at the public meeting, if it was, ultimately, not meaningful and potentially misleading.]
Lack of tools for dealing with BIAS
CNPP’s Obbagy described techniques her office will use to assess bias in studies. These tools can be seen on the USDA website here.
A disturbing observation is that Obbagy’s office is using a preliminary, untested tool for assessing bias in observational (epidemiological) studies—which are arguably most susceptible to bias, since these studies tend to have no pre-established protocols and can become virtual fishing expeditions for any desired result.
On Day 1, committee member Steven Hemsfield asked whether the CNPP had any procedures in place for dealing with bias in observational data. He said, “Earlier in my career I did a number of studies on self-reported dietary intake vs. actual intake, and I came to the conclusion that there’s pretty clear biases in that (self-reported) data.” (Day 1, Afternoon, 2:14:19)
Obbagy replied that this preliminary tool was the only one available. However, The Nutrition Coalition consulted academic epidemiologists who said that in fact, established tools do exist and are commonly used in the field. 
[UPDATE: A new tool for assessing bias in observational studies has been added to the website above. It is a tool that CNPP has developed—a modification of an existing tool—and does not appear to have been externally validated.]
An arguably far larger problem is a obvious pro-Guidelines bias within CNPP itself. Many CNPP staffers have worked on the Guidelines for a decade or more and have authored multiple publications supporting them, including, of course, publication of the Guidelines themselves. Thus, an observer could easily see an undeniable tendency among CNPP staffers to defend the existing approach. During the two-day meeting, there was no reckoning with the obvious question of why the Guidelines had thus far failed to stem the tremendous rise in obesity and diabetes, among other nutrition-related diseases. This was the clear elephant in the room.
To be clear, bias is natural. All researchers and institutions develop biases in favor of their preferred views of the science. Bias in nutrition science is clearly a serious problem and is one of the main reasons that most research findings in the field are false, according to Stanford professor John Ioannidis, who has documented this problem extensively and talks about in this lecture.
Lack of Prioritization of Clinical Trials Over Other Types of Data
International systems for scientific reviews always start by prioritizing clinical trials (which can demonstrate cause-and-effect) over observational studies (which can only show associations). However, Colette Rihane, a director at CNPP who’s been with the office for 15 years, stated that, contrary to these international standards, CNPP would not prioritize one type of science over another. “It should never be suggested that one approach is superior to another,” she emphasized in her review of types of scientific evidence (Day 1, afternoon, 7:15).
[UPDATE: CNPP states that Rihane was talking not about clinical trials vs. observational studies, which was her discussion at the moment of her comment, but rather the more general point that her office treats equally the three types of evidence it uses: 1. Systematic Reviews; 2. Data Analysis; and 3. Food Pattern Modeling. We have followed up with a question about whether CNPP will indeed prioritize clinical trials over observational data as a starting point for reviews.]
CNPP has historically not prioritized clinical trials as a superior form of evidence. A review of the office’s history shows a long track record of relying almost entirely on epidemiological evidence for its recommendations—resulting in failures and mistakes. Some examples are:
The citation of 34 epidemiological studies and only 1 clinical trial for a population-wide recommendation (in 2010) to eat a nutrient-dense breakfast for weight loss. More rigorous clinical trials later contradicted this advice, as reported in this Washington Post article: “How Government Nutritionists May Have Gotten it Wrong.”
Citation of mostly epidemiological studies for its low-fat dietary patterns while ignoring large, NIH-funded clinical trials on some 50,000 people contradicting this advice. Ultimately, this weak recommendation forced USDA-HHS to reverse itself and remove its formal “low-fat” diet advice. “There is no recommendation to reduce total fat intake or use fat free foods,” stated Alice Lichtenstein, Vice Chair of 2015 Guidelines.
The Nutrition Coalition believes that the longtime reliance of the Guidelines on this weak observational science to date is a major reason why this policy has failed to keep Americans healthy.
New Science Will be Ignored
(When CNPP Has Concluded That Science is Settled)
Julie Obbagy stated that if the CNPP staff had already deemed that the evidence for any food-disease relationship to be “strong,” her office would not revisit this topic, “even if new evidence comes out.” (Day 1, Afternoon. 30:34). This would apply to the supposedly “strong” evidence that the Guidelines’ high-carb diets are superior for weight loss/maintenance, a recommendation based on only one clinical trial conducted in Spain (Esposito 2004). This would also apply to the caps on saturated fats, even though meta-analyses and systematic reviews have universally concluded that these fats have no effect on cardiovascular or total mortality.
[UPDATE: CNPP reverses itself on this point: “USDA and HHS have not stated that a topic will not be examined if a previous systematic review has graded the evidence as strong.”]
CNPP May Not Update Existing Reviews, Despite Serious Known Flaws
CNPP says it will not necessarily update existing scientific reviews, even though the National Academies have found them lacking—i.e., not conducted according to state-of-the-art systematic methodology.  Whether old reviews will be updated will be determined by the committee, says Obbagy (Day 1, Afternoon, 2:26).
Specific problems with the CNPP’s existing reviews, according to an analysis in The BMJ were that the reviews ignored their own inclusion criteria and omitted the search dates as well as the names of the data bases used in some systematic reviews.
A fundamental component of reliability in scientific reviews is that they be transparent and reproducible. Because the CNPP reviews have sometimes fallen short of these standards, all the reviews should be scrutinized and updated where needed.
USDA Plans to Lean Heavily on Questionable Self-Reported Dietary Data
CNPP staff made clear that they plan to rely heavily on self-reported food-consumption data from the government’s “NHANES” data base, whose accuracy has been seriously questioned by more than one researcher and whose data have been called “implausible” and “invalid.” Continuing to use this data as the basis of our Dietary Guidelines is highly questionable.
Additional problems with NHANES were revealed in questions by committee members:
NHANES has no data on any of the brand foods items which comprise the majority of food items purchased at supermarkets (Day 2, afternoon 1:39) 
In the “Birth-to-24” category, for which the USDA-HHS are creating dietary recommendations for the first time, NHANES has data on a total of only 1,200 women Day 2, Morning, 1:02:09).
By the time of the 2020 Guidelines are published, the NHANES consumption numbers will already be a decade out of date (Day 2, afternoon, 1:40).
[UPDATE: CNPP states the following: “It is generally accepted in the nutrition field that self-report dietary data should not be used for drawing conclusions about energy balance and, in fact, it is not used in this way in the Dietary Guidelines process. However, it is generally accepted that self-report dietary data are valuable for providing important information at a population level regarding intakes (and sources) of foods and beverages and to describe dietary patterns and assess diet quality.” To read the full comments on this topic, see CNPP’s full response to our questions.]
USDA is Categorizing Foods in Non-Scientific Ways
Throughout the meeting, CNPP staff repeatedly lumped together “salt, sugar, and solid fats” in their analyses as if these were all nutrients to be equally minimized. It’s hard to understand this categorization, which has no basis in biology. Sugar, of course, is not necessary for health and can be minimized without harm. Yet more than a dozen studies now suggest that lowering salt excessively increases the risk of cardiovascular death.
Also, how are we to understand “solid fats,” which include both trans fats (the by-product of industrially produced vegetable oils) and also saturated fats (present in many natural foods eaten for millennia, such as meat, fish, dairy, and even olive oil). These two types of fats clearly do not belong in the same category.
Treating “salt, sugar, and solid fats” as one category is not scientifically justifiable.
CNPP also puts “calcium fortified soy foods” in the dairy category. Traditionally, dairy products are natural and come from ruminant animals. Clearly the soy industry has made inroads here, and it’s disappointing to see the USDA cater to this kind of industry influence.
CNPP Says A Nuanced Approach Towards Minority Groups is Possible—an Illusion?
Throughout the proceedings, committee members continually brought up the need for nuance in addressing the differing needs of minorities and underprivileged groups.
Committee members talked about the need to be sensitive to various cultural, ethnic and low-income subgroups (e.g., Katherine Dewey at min ~45:00 and Barbara Scheenman at 2:01:20 Day 1 afternoon).
Dr. Scheenman, chair of the advisory committee said, “Several people have referred to cultural aspects and yet we have foods reported by Americans…which kind of homogenizes us down to some very basic foods. I know that the Academy committee [National Academies of Sciences] was concerned about that, that we wind up with a generic food pattern as opposed to reflecting more cultural, ethnic patterns that could also…meet the Dietary Guidelines. I’m just wondering how far the committee can go to consider some of those issues.” (Continued from above, Day 1 afternoon)
After an evasive response from CNPP staff, Scheenman tries to follow up by asking, “Is it our job to think about if someone is coming with a Hispanic focus or an Asian focus, where would we do that?”
CNPP staff typically responded, telling committee members that they can “mention this in your report,” or “bring it up in your discussions in the report.”
For instance: On Day 2, USDA’s Janet de Jesus, tries to acknowledge concern by the expert committee that new scientific questions could be addressed, but says, “You’re welcome to discuss? But there won’t be an evidence-based review behind that….”
However, CNPP staff is aware that ultimately, the hundreds of pages written by the expert committee will be reduced to a single-page of guidelines, or even just a graphic like “MyPlate” Nuances and special considerations will largely be lost.
Why are USDA staffers not being more straightforward with committee members about the reality of this process?
On all the issues above, The Nutrition Coalition is submitting questions to CNPP to ensure that 1. we are understanding these issues correctly and 2. whether there is any additional, clarifying information that we could add. We will update this post with the responses.
Currently, it appears that if these critical process issues of the 2020 Guidelines are not fixed, there is little hope that these next guidelines will be based on a rigorous evaluation of the science.
To attend a future Guidelines meeting and make a public comment in person, see this blog post.
Note: CNPP’s website page for this first meeting does not have a transcript nor links to the videos of the proceedings. For your information, these are as follows:
Thursday, March 28: 9am-12:30pm
Thursday, March 28: 1:30pm-4:30pm
Friday, March 29: 8:30am-11:30am
Friday, March 29: 12:30PM-3:00pm
 The National Academies of Sciences, Engineering and Medicine, “Redesigning the Process for Establishing the Dietary Guidelines for Americans,” September, 2017, http://nationalacademies.org/hmd/Reports/2017/redesigning-the-process-for-establishing-the-dietary-guidelines-for-americans.aspx, 2-15.
 USDA staff said it collects this data but has not yet incorporated it into NHANES.
 Op. cit., S-4.